Faster Convergent Artificial Neural Networks
Journal Title: INTERNATIONAL JOURNAL OF COMPUTERS & TECHNOLOGY - Year 2018, Vol 17, Issue 1
Abstract
Proposed in this paper is a novel fast-convergence algorithm applied to neural networks (ANNs) with a learning rate based on the eigenvalues of the associated Hessian matrix of the input data. That is, the learning rate applied to the backpropagation algorithm changes dynamically with the input data used for training. The best choice of learning rate to converge to an accurate value quickly is derived. This newly proposed fast-convergence algorithm is applied to a traditional multilayer ANN architecture with feed-forward and backpropagation techniques. The proposed strategy is applied to various functions learned by the ANN through training. Learning curves obtained using calculated learning rates according to the novel method proposed are compared to learning curves utilizing an arbitrary learning rate to demonstrate the usefulness of this novel technique. This study shows that convergence to accurate values can be achieved much more quickly (a reduction in iterations by a factor of hundred) using the techniques proposed here. This approach is illustrated in this research work with derivations and pertinent examples to illustrate the method and learning curves obtained.
Authors and Affiliations
Dolores De Groff, Perambur Neelakanta
Keywords
Related Articles
- EP ID EP650962
- DOI 10.24297/ijct.v17i1.7106
- Views 91
- Downloads 0
WEB-BASED PENSION FUND MANAGEMENT SYSTEM
Given the discouraging record of the traditional pension system and the limited coverage of the private-sector pension schemes, there is need for robust web-based pension fundmanagement system that can handle the pension...
A MULTI AGENT DECISION TREE ALGORITHM FOR SUPPLY CHAIN MANAGEMENT
The issue of interest here is the behavior of decision agents in supply chain management. The study uses a managerial framework to organize supply-related tasks to the behavior of specific instantiations of decision agen...
Segmentation of Palmprint into Region of Interest (ROI): A Survey
Palmprint is one of the most reliable physiological characteristics that can be used to distinguish between individuals. Palmprint recognition process consists of image acquisition, pre-processing, feature extraction, ma...
A REVIEW OF NEW ABOUT CLOUD COMPUTING SECURITY
Job scheduling system problem is a core and challenging issue in Cloud Computing. How to use Cloud computing resources efficiently and gain the maximum profits with job scheduling. Job scheduling system is one of the Clo...
Recent innovations in Distributed Systems: Challenges and Benefits
The World Wide Web is used by millions of people everyday for various purposes including email, reading news, downloading music, online shopping or simply accessing information about anything. Using a standard web browse...